| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
Department of Biology, Washington University, St. Louis, Missouri 63130-4899 USA
Received for publication September 4, 2001. Accepted for publication January 10, 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: adaptive radiation • Asteraceae • Espeletia • Espeletiinae • Ichthyothere • ITS • Melampodiinae • paramo
Adaptive radiation in the Espeletiinae
The high Andes of northwestern South America contain a system of islandlike tropical montane habitat known as paramo. Although paramo has existed only since the final uplift of the Andes 2–4 million years ago (Van der Hammen and Cleef, 1986
), they are considered the most floristically diverse and endemic-species-rich high montane ecosystems in the world (Smith and Cleef, 1988
; Luteyn, 1999
). While many of the current paramo taxa are derived from ancestors in both northern and southern temperate mountains, others originated in the diverse surrounding tropical forests (Smith and Cleef, 1988
). Among the taxa thought to have evolved from ancestors in the tropical montane forests is the subtribe Espeletiinae (Asteraceae: Heliantheae). This group comprises eight genera and more than 100 described species endemic to Venezuela, Colombia, and Ecuador. The species are noted for their unusual giant rosette growth form and densely pubescent leaves, and many have come to dominate and characterize these tropical alpine landscapes.
Several characteristics of the Espeletiinae make the group ideal not only for gaining insight into the origin of paramo diversity, but also for understanding better the mechanisms of speciation and adaptive evolution in plants. Since their relatively recent origin in the late Pliocene or early Pleistocene, members of the Espeletiinae have shown a remarkable diversity of morphology, growth form, and ecological specialization (Cuatrecasas, 1986
). In addition to the giant polycarpic rosettes most characteristic of the group, there are tiny, mat-forming rosettes, sessile and caulescent monocarpic rosettes, branched shrubby trees, and even broad-leaved trees more than 10 m tall. Espeletiinae species can be found from upper montane cloud forests at about 2000-m elevation to the edge of glaciers at over 4600 m, having adapted to a wide range of conditions from gaps in the montane forest to wet paramo bogs to xeric periglacial talus slopes. Many have considered this rapid diversification a classic example of adaptive radiation in plants (Carlquist, 1974
; Monasterio and Sarmiento, 1991
).
Taxonomy and hypotheses of closest relatives
Despite much interest in the evolution, ecology, and ecophysiology of the Espeletiinae, phylogenetic relationships within the subtribe and between it and the rest of the Heliantheae are still unclear. The species assemblage was first described by Bonpland (von Humboldt and Bonpland, 1809
) as a single genus, Espeletia, with two species. Over the next two centuries, exploration of remote Andean mountaintops revealed dozens of new species. The morphological diversity and distinctiveness of these species convinced Cuatrecasas (1976)
that the creation of a new subtribe, Espeletiinae, was justified. Initially seven genera were recognized: Espeletia, Espeletiopsis, Coespeletia, Ruilopezia, Libanothamnus, Carramboa, and the monotypic genus Tamania. Later, an eighth genus, Paramiflos (also monotypic), was recognized (Cuatrecasas, 1995
).
While Bonpland initially believed that Espeletia sensu lato (s.l.) is closely related to the genus Silphium (von Humboldt and Bonpland, 1809
), Kunth (1820)
grouped it with genera such as Polymnia and Unxia. In one of the early monographs on Espeletia, Smith and Koch (1935)
identified Polymnia as likely the closest relative, although they gave few details supporting their hypothesis. Robinson (1981)
, however, determined that Polymnia, previously monographed by Wells (1965)
, is polyphyletic, comprising two phylogenetically distinct groups of species. Therefore, all but two species of Polymnia were removed from the genus and given the reestablished generic name Smallanthus, originally described by Mackenzie in 1933. Robinson (1981)
later commented on the similarities between the Espeletiinae and Smallanthus, thus implicitly favoring this genus as the closest relative to the Espeletiinae, rather than the revised Polymnia.
Additional candidates for the closest relatives of the Espeletiinae can be inferred from several tribal and subtribal taxonomic treatments. In the first publication that considered the Espeletiinae within a subtribal classification of the Heliantheae (Bentham and Hooker, 1873
), Espeletia s.l. was placed within the subtribe Melampodiinae, a system that continued until recently (for a detailed taxonomic history see Stuessy, 1973
, 1977
). In a more modern treatment of the Melampodiinae, Stuessy (1973)
created groups of generic affinity in which Espeletia s.l. was linked with both Polymnia (i.e., Smallanthus) and Unxia. Later, in his broader study of the Heliantheae, Stuessy (1977)
placed Espeletia s.l. in "group 2" of the Melampodiinae, which included Polymnia (Smallanthus), Rumfordia, Sigesbeckia, Trigonospermum, and Unxia; other genera of the subtribe were Acanthospermum, Lecocarpus, and Melampodium. While the Espeletiinae were recognized as a distinct subtribe by Robinson in his 1981 revision of tribal and subtribal limits in the Heliantheae, he thought that at least two other subtribes, the Melampodiinae and the Milleriinae, were closely allied. In this revision, the Melampodiinae were reduced from previous concepts of the subtribe to just five genera—Smallanthus, Acanthospermum, Ichthyothere, Lecocarpus, and Melampodium. The Milleriinae include several of the "group 2" Melampodiinae genera identified by Stuessy, such as Axiniphyllum, Rumfordia, and Sigesbeckia, along with Milleria and Guizotia.
More recently, Karis (1993)
applied modern cladistic techiniques to large sets of morphological characters across the entire tribe Heliantheae. In the results of this study, the Espeletiinae grouped not with taxa of the Melampodiinae, but rather in the subtribe Verbesiniinae, with genera such as Zaluzania, Encelia, Flourensia, Verbesina, and Perymenium. Both Smallanthus and Polymnia grouped with taxa traditionally associated with the Melampodiinae, such as Rumfordia, Melampodium, and Milleria. The most recent subtribal circumscriptions based on these data have therefore concluded that the Espeletiinae are neither deserving of subtribal status nor associated with the Melampodiinae (Karis and Ryding, 1994
).
Other research, however, has shown little support for this new systematic treatment. The first molecular study that included some of these taxa did not support the results of the morphological cladistic analysis. Using chloroplast restriction site data, Panero, Jansen, and Clevinger (1999)
found that species of the Espeletiinae were most closely related to the genera Smallanthus and Rumfordia in a majority of the most parsimonious trees. However, this study focused on the Ecliptiinae and therefore included only a few of the taxa in the Melampodiinae. Additional arguments for the Espeletiinae's association with the Melampodiinae are based on findings that at least some species of Espeletia contain melampolide sesquiterpene lactones (Torrenegra and Tellez A., 1995
), phytochemicals thought to be associated (though not exclusively) with taxa in the Melampodiinae.
Additional data are needed to resolve the phylogenetic position of the Espeletiinae within the Heliantheae and to elucidate the origins of this adaptive radiation. The internal transcribed spacer (ITS) region of nuclear ribosomal DNA has been an invaluable tool in the reconstruction of plant phylogenies (Baldwin et al., 1995
), especially at lower taxonomic levels, and has been useful in the identification of the closest relatives of other important adaptive radiations (e.g., Baldwin, 1992
). In this paper, I present the results of a molecular phylogenetic study of DNA sequence data from the nuclear ribosomal ITS region to address the question of which taxa are most closely related to the Espeletiinae.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
and Steussy (1977)
; in particular, taxa from the subtribes Melampodiinae and Milleriinae (sensu Robinson) were selected. Additional taxa (e.g., Zaluzania, Encelia) were included to test the alternative hypothesis suggested by the results of the morphological cladistic analysis (Karis, 1993
; Karis and Ryding, 1994
). After a preliminary analysis, several of the taxa that seemed closely related to the Espeletiinae were sampled more thoroughly to test for monophyly and to increase resolution on long branches. Within the Espeletiinae, taxa were chosen from all eight genera and included a broad sample of the morphological and geographical variation found in the subtribe.
DNA extraction, amplification, and sequencing
Total DNA was extracted from both silica-dried and herbarium material using a modified cetyltrimethylammonium bromide (CTAB) protocol (Doyle and Doyle, 1987
; Hillis et al., 1996
), with a 4% (mass per volume) CTAB solution. Three separate 20-µL polymerase chain reactions (PCRs) were set up for each individual in order to decrease the risk that observed polymorphisms would result from PCR error. Each reaction contained a final concentration of 2.5 mmol/L MgCl2, 10 mmol/L Tris-HCl (pH 9.0), 50 mmol/L KCl, 0.2 mmol/L of each deoxynucleotide, 0.1 mg/mL bovine serum albumine (BSA), 0.2 mmol/L of each primer, and 0.5 units/µL Taq polymerase. The cycling program included 30 s of denaturation at 94°C, followed by 35 cycles of amplification (94°C [20 s], 54°–56°C [35 s], 72°C [45 s]), and finally a single extension for 3 min at 72°C. Silica-dried samples were amplified as a single fragment using the primers ITS-1 (White et al., 1990
) (TCC GTA GGT GAA CCT GCG G) and ITS-4 (White et al., 1990
) (TCC TCC GCT TAT TGA TAT GC) at an annealing temperature of 54°C. ITS-1 and ITS-2 regions were amplified separately for herbarium samples using the following pairs of primers—(1) ITS-1 region: ITS-1 (White et al., 1990
) and ITS-1 5.8S (O'Kane, 1993
) (ACT CGA TGG TTC ACG GGA TT) with an annealing temperature of 54°C; (2) ITS-2 region: either ITS-3 (White et al., 1990
; as modified by O'Kane, 1993
) (GCA TCG ATG AAG AAC GTA GC) and ITS-4 (White et al., 1990
) (TCC TCC GCT TAT TGA TAT GC) with an annealing temperature of 54°C, or ITS-5.8F (AAG CCA YTA GGC CGA G) and ITS2–26S.4 (CGC CTG ACC TGG GGT CG) with an annealing temperature of 56°C.
The three reactions were combined into a single tube and visualized on a 1.8% agarose gel. Reactions that resulted in multiple bands (often representing amplification of fungal DNA associated with the leaves; J. T. Rauscher, unpublished data) were purified on a 1.8% agarose gel, and the bands of the appropriate size were excised. Excised bands and PCR solutions from single-band amplifications were purified using either a glass milk purification method (Hillis et al., 1996
) or Wizard PCR Preps (Promega, Madison, Wisconsin, USA). Most sequencing was done manually using the fmol DNA cycle sequencing system (Promega) labeled with [35S]
-dATP, run on a 6% Long Ranger (FMC BioProducts, Rockland, Maine, USA) gel, and exposed to x-ray film; other sequences were obtained using an automated sequencer (ABI 377; at the sequencing facility maintained by the International Center for Tropical Agriculture, Cali, Colombia) and ABI Prism BigDye Terminator Cycle Sequencing kits with Amplitaq DNA polymerase (Applied Biosystems, Foster City, California, USA). Both 5' and 3' strands were sequenced for each PCR product, and careful attention was paid to the potential for signal from more than one sequence (double bands or peaks), as might be expected in a multicopy gene such as ITS. GenBank accession numbers are listed in the Appendix (http://ajbsupp.botany.org/v89/).
Analysis
All sequences (including ITS-1, the 5.8S gene, and ITS-2) were aligned by eye using Se-Al, version 1.d1 (Rambaut, 1995
). Gaps were considered as missing data and either ignored or coded separately in a presence/absence matrix. Phylogenetic analysis was done using PAUP* 4.0b4 (Swofford, 2000
). Nucleotide sites with more than one band or peak were coded as degenerate sequence and treated as "polymorphic" under the parsimony options of PAUP*, because this variation most likely resulted from divergent repeat copies. Trees were constructed using the maximum parsimony criterion and an heuristic search with 1000 random additions (tree bisection-reconnection [TBR] branch swapping, MULPARS, ACCTRAN). From the resulting most parsimonious trees, a strict consensus tree was constructed. Clade support was estimated using both bootstrap values (500 searches with 10 random additions, MAXTREES = 100, implemented in PAUP*) and decay-index values (Bremer, 1988
; Donaghue et al., 1992
). In addition, pairwise sequence divergences were calculated for all taxa.
To examine alternative, less parsimonious topologies, I constructed constraint trees in MacClade 4.0 (Maddison and Maddison, 1992
) that forced the Espeletiinae to form sister-taxon relationships with clades not suggested by the most parsimonious trees. Heuristic searches, with 500 random additions, were performed in PAUP* to find the shortest trees consistent with these alternative topologies. To test the statistical significance of the difference in length between the topologically constrained trees and the optimal trees, I used PAUP* to perform a nonparametric Wilcoxon's signed-ranks test (Templeton, 1987
; also see Mason-Gamer and Kellogg, 1996
). A single tree from the set of most parsimonious trees was compared with one of the trees found during each constrained search, and the P value calculated. If the P value was close to significant, at either the one-tailed or two-tailed level, several other randomly chosen trees were tested to estimate the range of possible P values that might result from comparing different equally parsimonious trees within each of the sets of trees.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
; Robinson, 1981
). To identify the true root of the tree requires a much broader analysis of the entire tribe Heliantheae.
|
|
and Robinson, 1981
) or Milleriinae (sensu Robinson, 1981
) (Fig. 1, clades I–III). Of these taxa, the genus Ichthyothere is most closely related to the Espeletiinae, supported by a minimum of six unambiguous nucleotide changes, a decay-index value of two, and a bootstrap value of 73%. Other close relatives include the genus Smallanthus (clearly phylogenetically distant from the three species of Polymnia sampled in this study) and one species of Rumfordia. The sister taxon to this grouping (Fig. 1, clade I) is a clade including Trigonospermum, Milleria, Sigesbeckia, Axiniphyllum, and the other three sampled species of Rumfordia (Fig. 1, clade II). Somewhat more distantly related is a group including Melampodium, Acanthospermum, and Lecocarpus (Fig. 1, clade III).
In order to estimate the support of these data for grouping the Espeletiinae with clades other than just Ichthyothere, several constraint trees were created. The results for each of the alternative topologies tested, including the length of the most parsimonous constrained trees, the number of additional steps required, and the P value resulting from Templeton's (1987)
application of Wilcoxon's signed-ranks test are shown in Table 2. Among the suboptimal alternatives analyzed, the next most parsimonious reconstruction of the ITS data was a tree 1451 steps long (two steps less parsimonious than the optimal tree) in which Ichthyothere, Smallanthus, and Rumfordia guatemalensis (J.M. Coult) S.F. Blake form a clade that is the sister taxon to the Espeletiinae (Table 2b). A tree with four additional steps makes a clade containing just Ichthyothere and Smallanthus the sister taxon (Table 2c). Trees that force Smallanthus to be the closest relative of the Espeletiinae require five extra steps (Table 2d), as do three other topologies (Table 2e, f, g). Most of the alternative topologies listed in Table 2 are not significantly different from the most parsimonious tree under the Templeton test (P < 0.05, two-tailed). Topologies in which the Espeletiinae clade was forced outside of the entire clade of Melampodiinae and Milleriinae resulted in significant P values (Table 2h).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), there is no evidence for the relationship between the Espeletiinae and Verbesininae suggested by morphological cladistic analyses (Karis, 1993
); in fact, the Templeton test statistically rejects the possibility that the Espeletiinae fall outside the larger group of Melampodiinae and Milleriinae (Table 2h). These results make it difficult to justify subtribal status for the eight genera of the Espeletiinae. Although clearly a monophyletic group (see below), this clade is phylogenetically nested within the sampled taxa, and, as a result, formal taxonomic recognition of the Espeletiinae will make any broad concept of the Melampodiinae necessarily paraphyletic. Instead, it may be best to recognize these genera as a closely related complex within the Melampodiinae. For the remainder of this discussion, these genera will be collectively called the "Espeletia complex."
In general, the relationships suggested by the ITS phylogeny of these taxa support previous noncladistic circumscriptions of the Melampodiinae (or sections within it) based on morphological data. Two of the major clades resulting from this study (Fig. 1, clades I and II; clade III) correspond remarkably well to the "group 1" (Melampodium, Acanthospermum, and Lecocarpus) and "group 2" (Espeletia, Polymnia [Smallanthus], Rumfordia, Sigesbeckia, Trigonospermum, and Unxia [not sampled in this study]) Melampodiinae suggested by Stuessy (1977)
. While Robinson's concept of the subtribe Milleriinae seems to be an artificial subset of a broader Melampodiinae based on these data, his work (Robinson, 1981
) recognizes several taxa that Stuessy did not consider part of this group, such as Ichthyothere, Axiniphyllum, and Milleria.
Ichthyothere: sister taxon to the Espeletia complex?
This study is the first to suggest that species of the genus Ichthyothere are very closely related to and perhaps the closest extant relatives of the Espeletia complex. Although Robinson's previous taxonomic work suggested that Ichthyothere might be closely related (at least in a closely allied subtribe), it was never proposed as the closest relative. This is not surprising given the significant morphological evolution both groups have undergone since their divergence. The Espeletia complex's radiation of morphology and ecology is well known (Cuatrecasas, 1976
; Monasterio and Sarmiento, 1991
). Ichthyothere is also morphologically distinct compared with its closest relatives. Its characteristically small capitulum with few (usually one or two) fertile ray florets has contributed to its uncertain taxonomic position within the Heliantheae. In the most recent treatments, Ichthyothere has been placed in the Melampodiinae (Robinson, 1981
), the Milleriinae (Stuessy, 1977
), and a group of taxa unclassified to subtribe (Karis and Ryding, 1994
).
Despite this morphological divergence, there may be unique characters that support the relationship between the Espeletia complex and Ichthyothere. One such character is the lack of an expanded outer series of herbaceous involucral bracts. Most taxa within the Melampodiinae and Milleriinae (sensu Robinson) possess a distinct, expanded outer series of involucral bracts, or phyllaries. Smallanthus and Rumfordia, in particular, are noted for their broad, leafy phyllaries. In Ichthyothere, however, the phyllaries tend to be highly reduced. Similarly, the Espeletia complex lacks this distinct outer series, although the phyllaries are not reduced to the extent they are in Ichthyothere. A single exception in the Espeletia complex is the species Paramiflos glandulosa Cuatrec., which does possess a distinct outer series of glandular phyllaries (Cuatrecasas, 1995
); Paramiflos, however, is almost certainly nested within the Espeletia complex (Fig. 1) so this character is probably not plesiomorphic. Alone, the lack of expanded phyllaries is, at best, weak morphological evidence for the relationship between these two groups. More detailed study of the development and anatomy of the phyllaries and other morphological characters in these taxa is needed.
Alternative hypotheses
The ITS data alone provide only moderate support for the relationship between the Espeletia complex and Ichthyothere. Although the branch uniting these taxa includes six unambiguous nucleotide synapomorphies, homoplasy in these characters results in only moderate bootstrap (73%) and decay-index (2) values. Given the moderate support for this relationship, alternative phylogenetic hypotheses must be considered. The approach used in this study was to create an explicit hierarchy of likely alternative hypotheses and estimate their relative support given this data set. Relative support for each alternative can be demonstrated by both the number of additional steps that would be required on the tree and the approximate P value resulting from a Templeton test comparing the alternative trees with the most parsimonious trees.
This hierarchy of likely alternatives is shown in Table 2 by the number of additional steps required to force different sister taxon relationships. The next most parsimonious topology supported by the ITS data makes the Espeletia complex the sister taxon to a diverse clade containing Ichthyothere, Smallanthus, and Rumfordia guatemalensis; at a cost of only two additional steps (Table 2c), this should be considered an important alternative hypothesis. Though somewhat less likely given the additional five steps required, Smallanthus or R. guatemalensis may also be candidates for being closest relatives of the Espeletia complex. Because alternative topologies less than 11 or 12 steps longer than the most parsimonious tree cannot be rejected statistically using the Templeton test (Table 2), these less parsimonious alternatives should be considered until additional phylogenetic data are available.
Monophyly of the Espeletia complex
The ITS sequence data strongly support the monophyly of the Espeletia complex. This result is not surprising, given the number of cytological and morphological synapomorphies shared by members of this group. All species examined to date are exclusively diploid with a base chromosome number of n = 19 (Powell and King, 1969
; Powell and Cuatrecasas, 1970
, 1975
; Powell and Powell, 1978
; Spooner et al., 1995
; Carr et al., 1999
). Cuatrecasas (1986)
identified several morphological features that distinguish the Espeletia complex, such as spirally arranged xeromorphic or coriaceous leaves as well as aspects of achene structure and other floral characteristics. Molecular evidence supporting the monophyly of the complex has also been presented in chloroplast restriction studies (Panero, Jansen, and Clevinger, 1999
).
Because of the small number of taxa included in this analysis, a detailed discussion of relationships within the Espeletia complex will be reserved for forthcoming results. However, these preliminary analyses suggest a taxonomically relevant pattern worth mentioning briefly. Several of the genera proposed by Cuatrecasas (1976)
seem to be polyphyletic and/or paraphyletic on the ITS gene tree, including Espeletia, Espeletiopsis, Coespeletia, and Ruilopezia. Additional sampling within the complex, however, is required before the phylogenetic status of these genera can be reliably assessed.
Ichthyothere
The genus Ichthyothere is a group of about 25 species distributed in South and Central America. The monophyly of the genus was strongly supported by both bootstrap and decay-index values (Fig. 1) as well as by a single base-pair deletion (Fig. 2). Although sampling was limited in this study to only seven species (and only a single representative from each species), the resolution detected within Ichthyothere permits some preliminary discussion of evolution within the genus. Most notable is the well-supported basal split between I. scandens S.F. Blake and all other species sampled in this study. Ichthyothere scandens, found in Central and northwestern South America, differs from the rest of the genus by 14 unambiguous characters, including two deletions, one of which is a rare 5-bp deletion on the 5' end of the ITS-1 sequence. In the original publication of this species, Blake (1921)
described I. scandens as morphologically distinct from all other species of Ichthyothere on the basis of its scandent habit and its capitula grouped in loose panicles. Robinson (1980)
also recognized the distinctiveness of this species and suggested that it and another species from Colombia, I. garcia-barrigae H. Rob. (not sampled in this study), belong to a unique subgenus within Ichthyothere. Both are typically found at elevations above 1000 m.
The other species of Ichthyothere sampled here are small, erect herbs characterized by tightly clustered or glomerulate groups of capitula and are generally found below 1000 m elevation in or near the Amazon basin. The first of two clades found within this group of species includes I. granvillei H. Rob. and I. macdanielii H. Rob. Robinson's work (1983)
suggests that species such as I. davidsei H. Rob. and perhaps I. petiolata H. Rob. may be closely related to this clade as well. The second clade includes I. rufia Gardn., I. agrestis Baker, I. terminalis (Sprenge.) S.F. Blake, and I. latifolia Baker and shows little to no interspecific sequence variation.
Smallanthus
The genus Smallanthus, though perhaps not the sister taxon to the Espeletia complex, is clearly a close relative. Like the Espeletia complex and Ichthyothere, the monophyly of the genus, for the taxa sampled, is highly supported by both bootstrap and decay-index analyses (Fig. 1). Robinson's decision to separate Smallanthus from the genus Polymnia (Robinson, 1978
) is well justified; the three sampled species of Polymnia form a distinct clade, which is phylogenetically distant from Smallanthus and from all taxa traditionally included in or considered close to the Melampodiinae.
Among the 13 species of Smallanthus sampled in this study (from about 20 known species), the ITS tree contains at least four distinct lineages. The first includes five taxa—four South American species, S. jelskii (Hieron.) H. Rob., S. pyramidalis (Triana) H. Rob., S. sonchifolius (Poepp. & Endl.) H. Rob., and S. microcephalus (Hieron.) H. Rob., and one Mexican–Central American species, S. oaxacanus (Sch. Bip. ex Klatt) H. Rob., the sister taxon to the South American clade. This split between South American and Central American species is also found in two of the other major lineages. In the first, the South American species S. siegesbeckius (DC.) H. Rob. groups with the Central American species S. fruticosus (Benth.) H. Rob. and S. connatus (Spreng.) H. Rob.; in the second, S. meridensis (Steyerm.) H. Rob. and S. riparius (Kunth) H. Rob. (northwestern South America) group with S. maculatus (Cav.) H. Rob. and S. quichensis (Coult.) H. Rob. (Central America). That three of the major clades of Smallanthus contain both South American and Central or North American species suggests that dispersal across the Isthmus of Panama has occurred multiple times in the history of the genus. The fourth major clade contains a single species, S. uvedalius (L.) Mackenzie, the only species distributed as far north as the United States.
Rumfordia
On the basis of the results of this study, the genus Rumfordia is probably not monophyletic. Although monophyly cannot be rejected unequivocally by the Templeton test, the additional eight steps required to make Rumfordia monophyletic (Fig. 3e) suggest that it is unlikely. Rumfordia is either paraphyletic (both with respect to the Sigesbeckia-Axiniphyllum–Trigonospermum–Milleria clade and the Espeletia complex-Ichthyothere-Smallanthus clade) or polyphyletic in the most parsimonious trees (Fig. 3a, b); the specific topology is poorly resolved owing to a lack of characters that identify the proper position of the R. guatemalensis (Rumfordia I) clade. The suboptimal tree, in which Rumfordia is still paraphyletic but R. guatemalensis is basal with respect to the Trigonospermum-Milleria-Sigesbeckia-Axiniphyllum clade (Fig. 3c), is only two steps less parsimonious and thus a strong alternative hypothesis.
The molecular distinction between the R. guatemalensis and R. floribunda-revealii-penninervis (Rumfordia II) clades is not unexpected given previous morphological work. Both Greenman (1903)
and Sanders have noted that R. guatemalensis is more similar to species of Smallanthus based on its "habit, leaf form, texture and venation, inflorescence pubescence, phyllary shape, ray- and disk-corolla structure and number, and anther structure" (Sanders, 1977
, p. 302). Its placement in the genus Rumfordia, however, was justified mainly by its possession of perfect disc florets, thought to be a synapomorphy of the genus or a larger clade; other closely related taxa (e.g., the Espeletia complex, Smallanthus, Ichthyothere) have male-sterile disc florets. Additional evidence for the distinction between these two clades comes from chromosome numbers; R. floribunda has a base number of n = 24 (Sanders, 1977
) while R. guatemalensis is n = 13 (Strother, 1983
).
Other taxa and implications for disc-floret evolution
Closely related to the Rumfordia II clade is a well-supported (98% bootstrap, decay-index 10) clade of genera: Trigonospermum, Milleria, Sigesbeckia, and Axiniphyllum. This clade is also supported by chromosome counts, as all taxa counted to date have a base number of n = 15 or are polyploids from this base number (Turner and King, 1964
; Powell and King, 1969
; Solbrig et al., 1972
; Keil and Stuessy, 1975
, 1977
; Jansen et al., 1984
; Lane and Li, 1993
; Hu and Gu, 1996
). Both clades of Rumfordia as well as Sigesbeckia and Axiniphyllum are unusual with respect to the rest of the Melampodiinae because of their perfect disc florets; all other taxa have functionally male disc florets. Although variation in this character originally led Bentham and Hooker (1873)
to place taxa with perfect disc florets, such as Rumfordia and Sigesbeckia, outside the subtribe Melampodiinae, the close relationship between these species was presaged by several authors, including McVaugh and Anderson (1972)
, McVaugh and Laskowski (1972)
, Stuessy (1977)
, Turner (1978a
, b)
, and Robinson (1981)
.
If having functionally male disc florets is plesiomorphic for the taxa of interest in this study (as is suggested by the fact that all the closest outgroups, including Melampodium, Lecocarpus, Acanthospermum, and Guardiola, have this character state), then both of the two most parsimonious trees require a minimum of three changes: either a change to hermaphroditic disc florets and two reversals (Fig. 3a) or two changes to hermaphroditic discs and a single reversal to functionally male discs (Fig. 3b). In an alternative tree, two steps less parsimonious, in which Rumfordia is paraphyletic (Fig. 3c), only two changes are required, one change to perfect discs and a single reversal. While it is unclear whether the true history of these taxa involved two or three changes in disc-floret system of mating, it is certain that more than a single change is required, since a single change to hermaphroditic disc florets requires a tree 21 steps less parsimonious and statistically rejected under the Templeton test (Fig. 3f). It is not surprising that this character can be quite evolutionarily labile in these taxa and other Heliantheae. Stuessy (1973)
has previously discussed the artificiality of grouping genera by the sterility of disc florets while Turner (1978a)
has argued that frequent changes in disc-floret function between closely related species of Chrysanthellum resulted from a simple underlying genetic system.
As mentioned previously, the well-supported clade that contains Melampodium, Acanthospermum, and Lecocarpus (Fig. 1, clade III) agrees with the informal "group 1" Melampodiinae identified by Stuessy (1973)
on the basis of shared possession of inner involucral bracts that enclose and fuse with the ray achenes. While the ITS data in this study provide no information with respect to the monophyly of Melampodium, there is strong support for the relationship between continental species of Acanthospermum and the Galapagos Islands endemic genus Lecocarpus.
Chromosome evolution
The likely base chromosome number of several of the major lineages sampled here is well established; for others it is less clear. The Espeletia complex is n = 19, with no known polyploidy. The chromosomes of several species of Smallanthus have been counted (e.g., Turner, Powell, and King, 1962
; Solbrig et al., 1972
; Robinson et al., 1981
; Carr et al., 1999
), and although there is variation in the genus, including polyploid counts, the lowest and by far the most common count is n = 16, suggesting that it may be the base number. Chromosome number for Ichthyothere is known for only two species, one of which is n = 16 (Turner et al., 1979
), the other, a likely polyploid, with a number of ca. 33 (Coleman, 1970
). Therefore, estimating the base number is speculative at best. However, given that Smallanthus is probably a close relative and also has a count of n = 16, this base number may be accurate. As stated previously, the base number for the Rumfordia I clade is n = 13 while that for the Rumfordia II clade is n = 24 (possibly a result of polyploidy from an ancestor with a count of n = 12). A base number of n = 15 is likely for the clade including Axiniphyllum, Sigesbeckia, Trigonospermum, and Milleria. Finally, the base chromosome number for the clade containing Melampodium, Lecocarpus, and Acanthospermum is difficult to estimate at this time; species of Melampodium are quite variable, with low numbers ranging from 8 to 12 (e.g., Stuessy, 1972
; Keil and Stuessy, 1975
; Robinson et al., 1981
; Jansen et al., 1984
), while Lecocarpus and Acanthospermum both are n = 11 (Eliasson, 1970
; Jansen et al., 1984
; Husaini and Iwo, 1990
; Jose and Mathew, 1995
).
Figure 4 shows the estimated base chromosome numbers for the taxa of interest and a possible reconstruction of chromosome evolution inferred from the ITS gene tree. A hypothesis that can be reliably inferred from this tree is that the common ancestor of the Espeletia-Ichthyothere-Smallanthus clade had a base number of n = 16. The origin of the Espeletia complex may, therefore, be associated with an aneuploid increase in chromosome number to n = 19.
|
suggested that the ancestors of the complex were woody, a trait reflecting the supposed ancestral habit of the Heliantheae and Asteraceae. While the most recent common ancestor of the extant species of Espeletiinae may indeed have been a woody tree, there is no evidence that this is plesiomorphic. In fact, most species of Ichthyothere, Smallanthus, Rumfordia, Sigesbeckia, Axiniphyllum, Trigonospermum, and Milleria are herbs or erect herbs. Although some species of Smallanthus are woody subshrubs and small trees, there is no reason to assume that this habit is ancestral in this genus or any of the other closely related genera. Therefore, if the most recent common ancestor of the Espeletia complex was a woody tree, this habit was likely the result of relatively recent evolution from more herbaceous ancestors.
The time of origin of the Espeletia complex is not discussed in detail in this paper, but it is worth noting that levels of sequence divergence may be consistent with previously proposed hypotheses. Because the northern Andes uplifted to their present elevations only during the late Pliocene or early Pleistocene, the upper montane forests and paramos of this area are thought to have existed only for the last 2–4 million years (my; Van der Hammen and Cleef, 1986
). This approximate time range has been used as an estimate of when the early ancestors of the Espeletia complex arose. Similarly, in the silversword radiation of Hawaii, another adaptive radiation in the Heliantheae, the age of the most recent common ancestor has been estimated at about 5 my BP (Baldwin and Sanderson, 1998
). When the levels of sequence divergence between these two radiations are compared, they are quite similar. As shown in Table 1, the maximum nucleotide divergence within the Espeletia complex found in this study was approximately 3.9%; ITS studies in the silverswords reported a maximum within-group divergence of about 4.5%. In addition, the ca. 6–14% range of divergence between Espeletia and its purported closest relatives (Ichthyothere, Smallanthus and Rumfordia guatemalensis) is similar to that reported between the silverswords and their closest relatives in California, ca. 5.6–10.5%. Until we better understand the phylogeny of the Heliantheae, we cannot know how comparable the rates of sequence evolution between these two groups are. Nevertheless, assuming that these rates are not significantly dissimilar, the observed levels of sequence divergence may reflect the similar time scale in which evolution in these taxa is thought to have occurred.
Conclusions
From the consensus ITS gene tree presented in this analysis and its approximate concordance with previous taxonomic work, it is clear that the Espeletia complex is a member of the subtribe Melampodiinae and closely related to such genera as Ichthyothere, Smallanthus, and Rumfordia. While the precise intergeneric and interspecific relationships of these taxa and whether Ichthyothere is the closest extant relative of the complex are still not known, analysis of the most parsimonious and several less parsimonious trees has provided a more specific set of hypotheses of phylogeny, which can now be tested. Future work, including analysis of additional taxa and other nuclear and chloropast loci, should strengthen our phylogenetic inferences and broaden our insight into both evolution of the Melampodiinae and the origin and adaptive radiation of the Espeletia complex.
|
FOOTNOTES |
|---|
This work was presented to the Graduate School of Arts and Sciences of Washington University in partial fulfillment of the requirements for the degree of Doctor of Philosophy and was supported by a Doctoral Dissertation Improvement Grant from the National Science Foundation (grant no. 9623696). 
2 Current address: Liberty Hyde Bailey Hortorium, Department of Plant Biology, Cornell University, Ithaca, New York 14853 USA. E-mail for reprint requests: jtr24{at}cornell.edu
|
LITERATURE CITED |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Baldwin B. G. M. J. Sanderson 1998 Age and rate of diversification of the Hawaiian silversword alliance (Compositae). Proceedings of the National Academy of Sciences, USA 95: 9402-9406
Baldwin B. G. M. J. Sanderson J. M. Porter M. F. Wojciechowski C. S. Campbell M. J. Donoghue 1995 The ITS region of nuclear ribosomal DNA: a valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanical Garden 82: 247-277[CrossRef][ISI]
Bentham G. J. D. Hooker 1873 Melampodieae. In Genera plantarum, vol. 2, 190–192, 346–352. Reeve and Co., London, UK
Blake S. F. 1921 A remarkable new species of Ichthyothere. Journal of the Washington Academy of Sciences 11: 301-303
Bremer K. 1988 The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 795-803[CrossRef][ISI]
Carlquist S. 1974 Island biology. Columbia University Press, New York, New York, USA
Carr G. D. R. M. King A. M. Powell H. Robinson 1999 Chromosome numbers in Compositae. XVIII. American Journal of Botany 86: 1003-1013
Coleman J. R. 1970 Additional chromosome numbers in Brazilian Compositae. Rhodora 72: 94-99
Cuatrecasas J. 1976 A new subtribe in the Heliantheae (Compositae): Espeletiinae. Phytologia 35: 43-61
Cuatrecasas J. 1986 Speciation and radiation of the Espeletiinae in the Andes. In F. Vuilleumier and M. Monasterio [eds.], High altitude tropical biogeography, 267–303. Oxford University Press, New York, New York, USA
Cuatrecasas J. 1995 A new genus of the Compositae: Paramiflos (Espeletiinae) from Colombia. Proceedings of the Biological Society of Washington 108: 748-750[ISI]
Donaghue M. J. R. G. Olmstead J. F. Smith J. D. Palmer 1992 Phylogenetic relationships of Dipsacales based on rbcL sequences. Annals of the Missouri Botanical Garden 79: 249-265
Doyle J. J. J. L. Doyle 1987 A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11-15
Eliasson U. 1970 Studies in Galapagos plants. VIII. Chromosome numbers of some endemic species. Botaniska Notiser 123: 149-154[ISI]
Greenman J. M. 1903 New angiosperms from Mexico. Proceedings of the American Academy of Arts and Sciences 39: 99-100
Hillis D. M. B. K. Mable A. Larson S. K. Davis E. A. Zimmer 1996 Nucleic acids, IV: sequencing and cloning. In D. M. Hillis, C. Moritz, and B. K. Mable [eds.], Molecular systematics, 321–381. Sinauer, Sunderland, Massachusetts, USA
Hu X.-J. Z.-J. Gu 1996 A karyomorphological study of three species of genus Siegesbeckia. Acta Botanica Yunnanica 18: 77-80
Husaini S. W. H. G. A. Iwo 1990 Cytology of some weedy species of the family Compositae (Asteraceae) from Jos Plateau, Nigeria. Feddes Repertorium 101: 49-62
Jansen R. K. T. F. Stuessy S. Díaz-Piedrahíta V. A. Funk 1984 Recuentos cromosómicos en Compositae de Colombia. Caldasia 14: 7-20
Jose J. C. P. M. Mathew 1995 Chromosome numbers in the south Indian Heliantheae (Compositae). Compositae Newsletter 27: 7-10
Karis P. O. 1993 Heliantheae sensu lato (Asteraceae), clades and classification. Plant Systematics and Evolution 188: 139-195[ISI]
Karis P. O. O. Ryding 1994 Tribe Heliantheae. In K. Bremer [ed.], Asteraceae: cladistics and classification, 559–624. Timber Press, Portland, Oregon, USA
Keil D. J. T. F. Stuessy 1975 Chromosome counts of Compositae from the United States, Mexico and Guatemala. Rhodora 77: 171-195
Keil D. J. T. F. Stuessy 1977 Chromosome counts of Compositae from Mexico and the United States. American Journal of Botany 64: 791-798[CrossRef][ISI]
Kunth C. S. 1820 Compositae. In Nova genera et species plantarum, 279–283. Paris, France
Lane M. A. J. Li 1993 Documented chromosome numbers, 1993. 1. Chromosome number reports in Compositae with emphasis on tribe Astereae of the southwestern United States and Mexico. Sida 15: 539-546
Luteyn J. L. 1999 Paramos: a checklist of plant diversity, geographical distribution, and botanical literature. New York Botanical Garden Press, Bronx, New York, USA
Mackenzie K. K. 1933 Carduaceae 56. Smallanthus. In J. K. Small [ed.], Manual of the southeastern flora, 1307, 1406. New York, New York, USA. [Reprinted 1972, Hafner, New York, New York, USA.]
Maddison W. P. D. R. Maddison 1992 MacClade, v. 4.0. Sinauer, Sunderland, Massachusetts, USA
Mason-Gamer R. J. E. A. Kellogg 1996 Testing for phylogenetic conflict among molecular data sets in the tribe Triticeae (Gramineae). Systematic Biology 45: 524-545[CrossRef][ISI]
McVaugh R. C. Anderson 1972 North American counterparts of Sigesbeckia orientalis (Compositae). Contributions from the University of Michigan Herbarium 9: 485-493
McVaugh R. C. W. Laskowski 1972 The genus Trigonospermum Less. (Compositae, Heliantheae). Contributions from the University of Michigan Herbarium 9: 495-506
Monasterio M. L. Sarmiento 1991 Adaptive radiation of Espeletia in the cold Andean tropics. Trends in Ecology and Evolution 6: 387-391[CrossRef]
O'Kane S. L. 1993 Molecular systematics of Lopezia (Onagraceae). Ph.D. dissertation, Washington University, St. Louis, Missouri, USA
Panero J. L. R. K. Jansen J. A. Clevinger 1999 Phylogenetic relationships of subtribe Ecliptinae (Asteraceae: Heliantheae) based on chloroplast DNA restriction site data. American Journal of Botany 86: 413-427
Powell A. M. J. Cuatrecasas 1970 Chromosome numbers in Compositae: Colombian and Venezuelan species. Annals of the Missouri Botanical Garden 57: 374-379[CrossRef][ISI]
Powell A. M. J. Cuatrecasas 1975 IOPB chromosome number reports, L. Taxon 24: 671-678
Powell A. M. R. M. King 1969 Chromosome numbers in the Compositae: Colombian species. American Journal of Botany 56: 116-121[CrossRef][ISI]
Powell A. M. S. A. Powell 1978 Chromosomes numbers in Asteraceae. Madroño 25: 160-169
Rambaut A. 1995 Se-Al. (sequence alignment program). University of Oxford, Oxford, UK
Robinson H. 1978 Studies in the Heliantheae (Asteraceae). XII. Re-establishment of the genus Smallanthus. Phytologia 39: 47-53
Robinson H. 1980 Studies in the Heliantheae (Asteraceae). XXVI. New species of Ichthyothere. Phytologia 47: 128-134
Robinson H. 1981 A revision of the tribal and subtribal limits of the Heliantheae (Asteraceae). Smithsonian Contributions to Botany 51: 1-95
Robinson H. 1983 Studies in the Heliantheae (Asteraceae). XXX. A new species of Ichthyothere from Cayenne. Phytologia 53: 388-391
Robinson H. A. M. Powell R. M. King J. F. Weedin 1981 Chromosome numbers in Compositae. XII. Heliantheae. Smithsonian Contribributions to Botany 52: 1-28
Sanders R. W. 1977 Taxonomy of Rumfordia. Systematic Botany 2: 302-316[CrossRef][ISI]
Smith A. C. M. F. Koch 1935 The genus Espeletia: a study in phylogenetic taxonomy. Brittonia 1: 479-532[CrossRef]
Smith J. M. B. A. M. Cleef 1988 Composition and origin of the world's tropical alpine floras. Journal of Biogeography 15: 631-645[CrossRef][ISI]
Solbrig O. T. D. W. Kyhos M. Powell P. H. Raven 1972 Chomosome numbers in the Compositae. VIII. Heliantheae. American Journal of Botany 59: 869-878[CrossRef][ISI]
Spooner D. M. D. C. D. Dejong B. Y. Sun T. F. Stuessy K. M. Gengler G. L. Nesom P. E. Berry 1995 Chromosome counts of Compositae from Ecuador and Venezuela. Annals of the Missouri Botanical Garden 82: 596-602[CrossRef][ISI]
Strother J. L. 1983 More chromosome studies in Compositae. American Journal of Botany 70: 1217-1224[CrossRef][ISI]
Stuessy T. F. 1972 Revision of the genus Melampodium (Compositae: Heliantheae). Rhodora 74: 1-70,161-219
Stuessy T. F. 1973 A systematic review of the subtribe Melampodiinae (Compositae, Heliantheae). Contributions from the Gray Herbarium of Harvard University 203: 65-80
Stuessy T. F. 1977 Heliantheae: systematic review. In V. H. Heywood, J. B. Harborne, and B. L. Turner [eds.], The biology and chemistry of the Compositae, 621–671. Academic Press, London, UK
Swofford D. L. 2000 Phylogenetic analysis using parsimony (PAUP) version 4.0b4. Sinauer, Sunderland, Massachusetts, USA
Templeton A. R. 1987 Nonparametric phylogenetic inference from restriction cleavage sites. Molecular Biology and Evolution 4: 315-323[ISI][Medline]
Torrenegra R. D. A. N. Tellez A 1995 Chemotaxonomic value of melampolides in Espeletia species (Asteraceae). Biochemical Systematics and Ecology 23: 449-450[CrossRef]
Turner B. L. 1978a New species and combinations in the genera Sigesbeckia and Trigonospermum (Compositae: Melampodiinae). Brittonia 30: 64-68[CrossRef][ISI]
Turner B. L. 1978b Taxonomy of Axiniphyllum (Asteraceae-Heliantheae). Madroño 25: 46-52
Turner B. L. J. Bacon L. Urbatsh B. Simpson 1979 Chromosome numbers in South American Compositae. American Journal of Botany 66: 173-178[CrossRef][ISI]
Turner B. L. R. M. King 1964 Chromosome numbers in the Compositae. VIII. Mexican and Central American species. Southwestern Naturalist 9: 27-39[CrossRef]
Turner B. L. M. Powell R. M. King 1962 Chromosome numbers in the Compositae. VI. Additional Mexican and Guatemalan species. Rhodora 64: 251-271
Van der Hammen T. A. M. Cleef 1986 Development of the high Andean paramo flora and vegetation. In F. Vuilleumier and M. Monasterio [eds.], High altitude tropical biogeography, 153–201. Oxford University Press, New York, New York, USA
von Humboldt A. A. Bonpland 1809 Plantae aequinoctiales, vol. 2, 21–191. Paris, France
Wells J. R. 1965 A taxonomic study of Polymnia (Compositae). Brittonia 17: 144-159[CrossRef]
White T. J. T. Bruns S. Lee J. Taylor 1990 Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenies. In M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White [eds.], PCR protocols, 315–322. Academic Press, San Diego, California, USA
This article has been cited by other articles:
|
|
 |
|
  R. A. Scherson, R. Vidal, and M. J. Sanderson Phylogeny, biogeography, and rates of diversification of New World Astragalus (Leguminosae) with an emphasis on South American radiations Am. J. Botany, August 1, 2008; 95(8): 1030 - 1039. [Abstract] [Full Text] [PDF]
|
|
|
|
